|Publication number||US5889934 A|
|Application number||US 08/805,144|
|Publication date||Mar 30, 1999|
|Filing date||Feb 24, 1997|
|Priority date||Feb 24, 1997|
|Also published as||CA2275877A1, DE69727083D1, DE69727083T2, EP1019820A1, EP1019820B1, WO1998037491A1|
|Publication number||08805144, 805144, US 5889934 A, US 5889934A, US-A-5889934, US5889934 A, US5889934A|
|Inventors||Gary S. Peterson|
|Original Assignee||Data General Corporation|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (8), Non-Patent Citations (1), Referenced by (83), Classifications (9), Legal Events (4)|
|External Links: USPTO, USPTO Assignment, Espacenet|
This invention relates to high availability disk arrays or use in data processing systems and, more particularly, to improved techniques for handling various types of power failure conditions that may occur during operation of the system.
One particular disk array system which provides a high degree of availability of the disks thereof is often referred to as a Redundant Array of Inexpensive Disks (RAID). Such system uses an intelligent input-output (I/O) processor for accessing one or more data storage disks of the array in response to the needs of a host computer, each data storage disk of the array being driven by disk drive circuitry operating via the I/O control processor in a manner that effectively appears to the host computer as a single disk drive. A data storage disk module comprises, for example, a disk, disk driver circuitry, and power/control circuitry. Alternatively, in some implementations of such systems, an I/O processor need not be used and the host computer may communicate directly with the disk modules which form an array.
In a particular RAID-5 context, for example, which comprises an array of five data storage disks, each disk has a plurality of data storage sectors, corresponding sectors in each of the five disks being referred to as a sector group or "stripe" of sectors. Each stripe includes one sector for holding redundant, or parity, data. The remaining sectors in the stripe store user data. The use of such redundancy allows for the reconstruction of user data in the event of a failure of a user data sector in the stripe.
When a user data disk module fails, the redundant or parity entry that is available in the parity sector of a stripe and the data in the non-failed user data sectors of the stripe can be used to permit the user data that was in the sector of the failed disk to be effectively reconstructed so that the system can remain operative using such reconstructed data even when the user data of that sector of the failed disk cannot be accessed.
Certain kinds of failures, however, can occur in which the array is left in an incoherent or effectively unusable state, e.g., a situation can occur in which there is a power failure, i.e., power to the I/O processor (IOP) fails or the I/O processor itself fails due to a hardware defect, or power to the disk drives themselves fails. A further problem can arise, for example, if a power failure results in the need to use a new IOP to replace a failed one and there is no way to identify where a write operation to a sector of the array was taking place after the new IOP has replaced the old IOP.
Techniques have been devised for handling such power failure situations that cannot be handled by RAID-5 systems as originally designed. Using conventional disk drives to perform a RAID-5 style write typically requires several distinct commands and associated data transfers. A RAID-5 write begins with a command to read old data from the target data drive. Then, new data is written to the data drive. A partial product is obtained by XORing the old and new data to get a partial product. A read command is issued to read the old parity from the disk drive containing the parity sector. The partial product is XORed with the old parity. Finally, the result of the XOR is written into the parity sector. If a write is interrupted by a failure, a mechanism is needed to identify where an interrupted write may have been started.
In order to handle this situation, Data General Corporation, the assignee of the present application, has used, in prior art systems, additional bytes at the end of a sector to include what have been called a checksum, a shed stamp, a time stamp and a write stamp. Typically a sector includes 512 bytes of host data and 8 bytes of validation data. The checksum verifies that the host data is correct and is on the correct disk sector. The shed stamp is a series of bits used to identify whether the parity sector of a particular sector group contains parity data or shed data. Shed data is used when the array is being operated in a degraded mode, in other words when one of the disk drives is not operating. Shed data is described in U.S. Pat. No. 5,305,326 (Solomon et al.), the disclosure of which is hereby incorporated by reference herein.
A time stamp is used to provide a check against data sector corruption from the last major update. Each time there is a major stripe update (i.e., a write command that causes all sectors in a stripe to be updated), the time stamp is set in each of the sectors that belongs to the particular stripe. The time stamp is a unique random number that is written into each of the sectors in the stripe at the time of the major stripe update. Thereafter, any change to the data in that sector causes the time stamp to be invalidated. The time stamp, thus, provides a validation mechanism for sectors that have not been updated since the last major stripe update.
The write stamp is a series of bits, one for each data storage disk in a sector group. During a major stripe update, the write stamps are all set to zero. Each time a write is performed, the bit corresponding to the disk drive being updated is flipped in the drive being updated. The corresponding bit is also flipped in the parity drive upon completion of the write. In order to flip the write stamp bit corresponding to the disk drive being written into, a pre-read must take place so as to know the old value. Only then is it known what the new value should be. The pre-read is performed anyway during a RAID-5 write into a conventional disk drive in order to retrieve the partial product for writing the parity sector. The write stamp is used to provide a validation mechanism for sectors that have been updated since the last major stripe update.
Data storage disk drives with the ability to coordinate parity updates through drive-based XORing will soon be available on the market. When such a disk drive is operated in the drive-based XOR mode, a host simply issues a XOR-WRITE command to the disk drive. This command causes the data drive to read the old data from the disk, write the new data to the disk, compute the partial product, issue a partial product XOR-WRITE to the parity drive and then return status to the host when the entire update is complete. Thus, the host only has to issue a single command. Parity is updated by the disk drives and there is no need for any read command. To implement the prior art write stamp validation mechanism on a disk drive performing drive-based XOR would result in slowing the system operation. Thus, it has become desirable to develop a new validation system that does not involve pre-reads as was the case for the write stamp.
In accordance with the invention, a stripe, or sector group, includes a plurality of user data sectors and a parity sector. Each of the sectors in the group lies on a different disk drive in a disk array. Each of the sectors in the group includes a plurality of validation fields. Each of the validation fields on a sector has corresponding fields on the remaining sectors in the group. Each of the user data sectors is assigned to at least two selected validation fields for placement of its validation stamp. A validation stamp in a presently preferred embodiment is a unique number randomly generated at the time of a write to the respective disk drive. The fields for each user data sector are assigned so that no two user data sectors in the same group have their respective validation stamps in the same selected validation fields. While a single overlap of fields is permitted, no two user data sectors have their validation stamps in the same set of selected validation fields. The parity sector would include a parity entry in each of the validation fields that corresponds to a combination of the corresponding validation fields. In the case where drive-based XORs are used, the parity entry for any given validation field would be the XOR of all of the corresponding validation fields in the user data sectors of the group.
If one of the disk drives in the sector group becomes inoperative, the validation fields to which that disk drive was assigned become ineffective. However, since none of the other disk drives use those selected validation fields exclusively for their validation stamps, there will always be for each disk drive a validation stamp that is present in an operative validation field. Thus, the validation method of the present invention operates when the sector group is in a degraded mode as well as in normal full operation.
Other objects and advantages of the invention will become apparent during the following description of the presently preferred embodiment of the invention taken in conjunction with the drawings.
FIG. 1 is a diagrammatic map of how data and validation bits are arranged on a sector group in accordance with the present invention.
FIG. 2 is a diagrammatic map of how the validation stamp assignments are made in accordance with the present invention.
FIG. 3 is a representation of how the validation stamp arrangement of FIG. 2 handles degraded mode with a disk missing.
The present invention is implemented on a disk array. A plurality of data storage disks are related to one another by the stripes or sector groups that are assigned across the plurality of disks. Each sector group has a plurality of user data storage sectors and a single parity sector. Each sector in a group is located on a different one of the data storage disks. In accordance with a RAID-5 sector group, the parity sector contains the XOR of all the user data sectors in the sector group. The XOR may result from the conventional three step process of writing data, pre-reading and writing parity or from the one step write made possible with a drive-based XOR operation. The array of data storage disks is accessed by an I/O processor. The design of an I/O processor for controlling a disk array in accordance with a RAID algorithm is well within the skill of those in the pertinent field of computing.
Referring now to FIG. 1, the arrangement of data in the disk array in accordance with teachings of the present invention shall now be described. A sector group of five sectors located on five data storage disks is shown in FIG. 1. However, the present invention is not limited to a particular size of a sector group. For instance, a sector group may extend over eleven disk drives including one parity sector and ten data sectors. In FIG. 1, four data storage sectors are shown which in addition to one parity sector make up the sector group. Each of the sectors is located on a different data storage disk. Each data storage disk has a large number of sectors stored thereon. In order to describe the present invention, it is only necessary to illustrate a single sector group.
Each of the user data sectors includes a user data field which in accordance with the presently preferred embodiment is 512 bytes. This is the typical sector size employed in conventional data storage devices. Data from the host computer is stored in the user data sectors. One of the sectors in the group is a parity sector and provides parity for the user data sectors. If one of the data storage sectors should fail, the parity sector may be used in conjunction with the working data storage disks to reconstruct the user data from the nonfunctioning data storage device. In accordance with a RAID-5 algorithm, the parity sector is formed by XORing the user data sectors of the remaining user sectors in the sector group.
Additional data is added onto each sector to provide for handling data under power failure and other failure conditions. In accordance with the presently preferred embodiment, a checksum field is provided in addition to the user data. A parity invalid field may also be provided on each of the disk drives. The parity invalid field need only be a single bit. The parity invalid bit would always be zero in the user data sectors and thus, always zero in the parity drive when all of those bits are XORed. The parity invalid bit is only set on the parity drive by a direct write. If a partial RAID-5 update is detected on an active drive during a degraded verification of the sector group, the bit is set in the parity sector to indicate that the data is invalid. The data in the parity sector then will not be used. If the parity invalid bit is not set, the data checksums will still be checked for data validity.
The presently preferred embodiment further includes a shed stamp field to indicate whether the parity sector contains parity data or shed data from a data storage disk that has become inoperative. When a data storage device fails, parity data is used to reconstruct the data of the failed data storage disk. The reconstructed data can be stored in the parity disk drive replacing the parity data. Setting the shed stamp indicates that the disk drive contains the shed data from the failed disk rather than the normal parity data. When one of the data storage disks has failed in a sector group, the sector group is operating in a degraded mode and parity can no longer be maintained. Therefore, one has the option of maintaining the parity data and reconstructing the user data of the failed disk each time it is needed or replacing the parity data with reconstructed shed data so that user data from the failed disk can be accessed directly from the parity drive.
In accordance with the present invention, a validation stamp area is provided with a number of validation fields. The validation stamp area performs the prior art functions of the time stamp and the write stamp. As the validation fields of each disk drive are updated and the checksums are updated, the parity sector receives the XOR of all the validation areas and checksums in the sector group. The present invention may be used with data storage disks that perform a drive-based XOR to update the parity drive.
Referring now to FIG. 2, the validation area of the present invention is described in greater detail. FIG. 2 shows a validation stamp area that has been divided into five validation stamp fields. The arrangement shown in FIG. 2 is designed for a system that permits as many as ten data storage disk drives and one parity drive to be contained within a single sector group. In accordance with this presently preferred embodiment, each data storage disk drive is assigned two of the validation stamp fields for storing its validation stamp. The invention can just as easily be implemented by assigning three validation fields to each data storage disk drive. For example, by switching the unassigned fields to assigned fields and vice versa, a working three field assignment arrangement is achieved. For the invention to provide an efficient validation area, each data storage disk drive must be assigned at least two validation fields for its validation stamp and the number of validation fields is smaller than the number of user data sectors permitted in a sector group. The assignment of validation fields is arranged so that no two user data sectors are assigned to the same set of corresponding validation fields.
The validation stamps are further distributed so that each of the validation fields has roughly about the same number of validation stamps throughout the sector group. In FIG. 2, each of the validation fields has four validation stamps. The invention works as well if, for instance, data drive No. 9 were not included and two of the validation stamp fields therefore included only three validation stamps as shown in the figure.
While it is desirable to use as small a number of validation fields as possible to minimize the amount of memory lost to the validation technique, there is also a minimum number of validation fields required to carry out the invention. For example, when two validation fields are to be assigned to each disk drive for its validation stamp, there must be at least m+2 fields in total; where m is the largest value of m for which the summation of from 1 to m is less than the number of data drives. Thus, for a system with ten data disk drives and one parity drive, five validation fields are required. The sum of 1 plus 2 plus 3 is less than ten, whereas 1 plus 2 plus 3 plus 4 is not. The minimum number of required validation fields is 3 plus 2, because 3 is the largest value of m satisfying the algorithm.
The validation stamp is a random value assigned to a drive at the time of a write. During a major stripe update, each disk drive is assigned a random number for its validation stamp by the disk array controller firmware. The validation stamp of each drive is placed in its assigned validation fields. The remaining validation fields for that drive remain set to zero. The validation stamp of the embodiment shown for FIG. 2 is a 12-bit random value. After the major stripe update, as disk drives are updated through RAID-5 writes, each data drive would be rewritten with a new random value for its validation stamp which are placed into the assigned validation fields for that drive. As a user data sector is updated in this manner, the parity drive is also updated by the XORing of the data across the entire sector group.
The value of the validation stamp area of the present invention can be seen with respect to FIG. 3 which depicts what happens when one of the disk drives fails. The validation area of the invention allows the complete validation of all data storage disk drives even if one of the disk drives is missing at the time of validation. It is noted that when the data drive 3 is removed from operation in FIG. 3, two of the validation stamp fields and the fields corresponding thereto cannot be validated. Those include all validation fields corresponding to the fields containing the validation stamp for the failed disk drive. These fields cannot be validated because all of the data is not present. However, by providing every disk drive with a second copy of its validation stamp in a validation field that does not correspond to those assigned to the failed data drive, all the remaining drives can still be validated. Each of the remaining data drives has at least one validation field that can be validated. No two disk drives use two validation fields that correspond identically to two validation fields assigned to another disk drive in the sector group. Thus, each disk drive has a validation stamp in a field unaffected by the loss of one of the disk drives.
The validation area advantageously provides a stripe coherency check for the data drives even when in a degraded mode with a failed disk drive.
This validation stamp method will only fail to detect a partial write if coincidentally the new random number that is written into the validation stamp when a partial write occurs is the same as the old random number that was written into the validation stamp. For the 12-bit validation stamps, this works out to 99.976% data coverage. The enhanced reliability offered by the validation stamp is obtained at the expense of the extra data storage required by the validation area. The validation area and other checking mechanisms implemented in the presently preferred embodiment require a disk sector size of 524 bytes.
The validation system comes especially into play during the performance of certain operations that shall now be described. Verify operations are used to determine whether all data in the user data sectors of a sector group are valid. With all data drives present, a verify operation proceeds by locking the sector group (i.e., obtaining exclusive access to the group) so that write operations do not interfere with the data while the verify is ongoing. A command is then sent which asks for the XOR of all the data drives and the parity drive in the sector group. If the data is valid, the result of XORing the data drives and the parity drives should be zero in all cases. If the shed stamp is non-zero, that means the parity drive contains shed data rather than parity data and verification cannot actually take place. However, the shed data can be read and written into the data drive that has been returned to operation. The parity sector can then be rebuilt and the sector group is thereafter ready for verification. If the validation area, the checksum or any of the data bytes results in a non-zero result from this command, the parity drive is corrected to include the correct parity data. The sector group may then be unlocked. The verification operation makes sure that the RAID algorithm is fully being carried out into the parity sector.
The verify operation operates a little differently in the degraded mode where there is a single data drive missing. The sector group is locked. A command is issued to ask for an XOR of all the user data sectors and the parity sector in the sector group. In the validation field of the result buffer, the bits are zeroed out for those fields that correspond to those assigned to the missing disk drive. If the shed stamp is non-zero, that is how it should be, i.e., the parity sector has been converted to contain shed data from the missing drive instead of the normal parity data. If in the resultant data, the shed stamp is zero and any of the validation area, checksum or data bytes are non-zero, then a RAID-5 write was interrupted on a data drive. The lost parity cannot be regenerated because of the missing disk drive. At this point, the parity invalid bit is set on the parity drive with a direct write command. The missing disk drive's data is now unavailable and cannot be reconstructed by the data storage system. The sector group is then unlocked. The sector group is then available in its degraded mode without parity. It is up to the host whether it chooses to continue to use the sector group in that condition.
A rebuild operation is generally performed when a new data drive is returned to the sector group. The sector group is locked. The parity sector is read. If the parity invalid bit is set, the new data drive is filled with zeroes and an invalid checksum. The parity invalid bit indicates that valid data is not available for the new disk. Then the rebuild command is further issued to the parity drive to have the parity data rebuilt. If the shed stamp is set, then the new data drive would be written with the shed data from the parity drive. Then the rebuild command is further issued to the parity drive to have the parity data rebuilt. If neither the parity invalid nor the shed stamp is set, the new data drive's data is rebuilt using the parity sector and all of the active disk drive's user data. If desired, a verification operation can take place before unlocking the sector group.
It should be understood that various changes and modifications to the embodiments described above will be apparent to those skilled in the art. For example, a validation stamp may be any number of bits. The number of validation fields used may depend on the number of disk drives permitted in a sector group. Also, the number of validation stamps used per disk drive may depend on the number of disk drives permitted in a sector group. These and other changes can be made without departing from the spirit and the scope of the invention and without diminishing its attendant advantages. It is, therefore, intended that such changes and modifications be covered by the following claims.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US5195100 *||Mar 2, 1990||Mar 16, 1993||Micro Technology, Inc.||Non-volatile memory storage of write operation identifier in data sotrage device|
|US5305326 *||Mar 6, 1992||Apr 19, 1994||Data General Corporation||High availability disk arrays|
|US5463765 *||Mar 18, 1993||Oct 31, 1995||Hitachi, Ltd.||Disk array system, data writing method thereof, and fault recovering method|
|US5537567 *||Mar 14, 1994||Jul 16, 1996||International Business Machines Corporation||Parity block configuration in an array of storage devices|
|US5640506 *||Feb 15, 1995||Jun 17, 1997||Mti Technology Corporation||Integrity protection for parity calculation for raid parity cache|
|US5644767 *||Feb 7, 1995||Jul 1, 1997||At&T Global Information Solutions Company||Method and apparatus for determining and maintaining drive status from codes written to disk drives of an arrayed storage subsystem|
|US5737344 *||May 25, 1995||Apr 7, 1998||International Business Machines Corporation||Digital data storage with increased robustness against data loss|
|EP0662660A1 *||Dec 6, 1994||Jul 12, 1995||International Business Machines Corporation||An improved data storage device and method of operation|
|1||*||Ng et al., Uniform Parity Group Distribution in Disk Arrays with Multiple Failures, IEEE transactions on Computers, vol. 43, No. 4, Apr., 1994.|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US6167485 *||Dec 31, 1997||Dec 26, 2000||Emc Corporation||On-line data verification and repair in redundant storage systems|
|US6195761 *||Dec 31, 1997||Feb 27, 2001||Emc Corporation||Method and apparatus for identifying and repairing mismatched data|
|US6233696 *||Dec 31, 1997||May 15, 2001||Emc Corporation||Data verification and repair in redundant storage systems|
|US6243827 *||Jun 30, 1998||Jun 5, 2001||Digi-Data Corporation||Multiple-channel failure detection in raid systems|
|US6389414 *||Sep 21, 1998||May 14, 2002||Microsoft Corporation||Internal database validation|
|US6389511||Aug 9, 2000||May 14, 2002||Emc Corporation||On-line data verification and repair in redundant storage system|
|US6535994||Jul 5, 2000||Mar 18, 2003||Emc Corporation||Method and apparatus for identifying and repairing mismatched data|
|US6560598||Oct 16, 2001||May 6, 2003||Microsoft Corporation||Internal database validation|
|US6584544 *||Jul 12, 2000||Jun 24, 2003||Emc Corporation||Method and apparatus for preparing a disk for use in a disk array|
|US6687850 *||Jan 31, 2001||Feb 3, 2004||Western Digital Technologies, Inc.||Disk drive for storing sector-reconstruction sectors and for storing a sector-reconstruction status in sectors distributed around a disk|
|US6785835 *||Jan 25, 2001||Aug 31, 2004||Hewlett-Packard Development Company, L.P.||Raid memory|
|US6865650 *||Sep 29, 2000||Mar 8, 2005||Emc Corporation||System and method for hierarchical data storage|
|US6950900||Sep 27, 2000||Sep 27, 2005||International Business Machines Corporation||Method and apparatus for migrating data having a format of a first type to a format of a second type|
|US6981102 *||Oct 11, 2002||Dec 27, 2005||International Business Machines Corporation||Method and system for managing meta data|
|US7020805||Aug 15, 2002||Mar 28, 2006||Sun Microsystems, Inc.||Efficient mechanisms for detecting phantom write errors|
|US7036066||May 24, 2002||Apr 25, 2006||Sun Microsystems, Inc.||Error detection using data block mapping|
|US7051155||Aug 5, 2002||May 23, 2006||Sun Microsystems, Inc.||Method and system for striping data to accommodate integrity metadata|
|US7103811||Dec 23, 2002||Sep 5, 2006||Sun Microsystems, Inc||Mechanisms for detecting silent errors in streaming media devices|
|US7133883||Dec 23, 2002||Nov 7, 2006||Sun Microsystems, Inc.||General techniques for diagnosing data corruptions|
|US7173782 *||Mar 17, 2005||Feb 6, 2007||Hitachi Global Storage Technologies Netherlands, B.V.||Magnetic disk drive and refresh method|
|US7353432||Nov 25, 2003||Apr 1, 2008||Sun Microsystems, Inc.||Maintaining high data integrity|
|US7467169||Oct 31, 2005||Dec 16, 2008||Network Appliance, Inc.||Circular and bi-directional mirroring of flexible volumes|
|US7487400 *||Aug 29, 2006||Feb 3, 2009||Hitachi, Ltd.||Method for data protection in disk array systems|
|US7552357||Apr 29, 2005||Jun 23, 2009||Network Appliance, Inc.||Lost writes detection in a redundancy group based on RAID with multiple parity|
|US7577804 *||Oct 6, 2004||Aug 18, 2009||International Business Machines Corporation||Detecting data integrity|
|US7590798||Dec 15, 2003||Sep 15, 2009||Netapp, Inc.||Method and system for responding to file system requests|
|US7716420||Apr 28, 2006||May 11, 2010||Network Appliance, Inc.||Methods of converting traditional volumes into flexible volumes|
|US7752180||Jul 6, 2010||Network Appliance, Inc.||File system group consistency point|
|US7805568||Sep 28, 2010||Spinnaker Networks, Llc||Method and apparatus for data storage using striping specification identification|
|US7809868||Apr 23, 2007||Oct 5, 2010||Network Appliance, Inc.||System and method for filtering information in a data storage system|
|US7844701||Nov 30, 2010||Network Appliance, Inc.||Rule-based performance analysis of storage appliances|
|US7856530||Dec 21, 2010||Network Appliance, Inc.||System and method for implementing a dynamic cache for a data storage system|
|US7870317||Sep 19, 2005||Jan 11, 2011||Network Appliance, Inc.||Storage processor for handling disparate requests to transmit in a storage appliance|
|US7877543 *||Jan 25, 2011||Hewlett-Packard Development Company, L.P.||System and method for writing data and a time value to an addressable unit of a removable storage medium|
|US7908656||Mar 15, 2011||Network Appliance, Inc.||Customized data generating data storage system filter for data security|
|US7917693||Mar 29, 2011||Netapp, Inc.||Method and system for responding to file system requests|
|US7945724||May 17, 2011||Netapp, Inc.||Non-volatile solid-state memory based adaptive playlist for storage system initialization operations|
|US7996606 *||Dec 13, 2010||Aug 9, 2011||Hewlett-Packard Development Company, L.P.||System and method for writing data and a time value to an addressable unit of a removable storage medium|
|US8032697||Jun 26, 2009||Oct 4, 2011||Netapp, Inc.||Method and system for responding to file system requests|
|US8078622||Oct 30, 2008||Dec 13, 2011||Network Appliance, Inc.||Remote volume access and migration via a clustered server namespace|
|US8131926||Oct 20, 2004||Mar 6, 2012||Seagate Technology, Llc||Generic storage container for allocating multiple data formats|
|US8156241||May 17, 2007||Apr 10, 2012||Netapp, Inc.||System and method for compressing data transferred over a network for storage purposes|
|US8166257||Apr 24, 2012||Network Appliance, Inc.||Automated continuous provisioning of a data storage system|
|US8176012||May 8, 2012||Netapp, Inc.||Read-only mirroring for load sharing|
|US8214589||Mar 13, 2009||Jul 3, 2012||Seagate Technology Llc||Data storage system redundancy scheme verification|
|US8219887 *||Jul 10, 2012||Marvell World Trade Ltd.||Parallel Reed-Solomon RAID (RS-RAID) architecture, device, and method|
|US8321867||Nov 27, 2012||Network Appliance, Inc.||Request processing for stateless conformance engine|
|US8359524||Jun 29, 2012||Jan 22, 2013||Marvell World Trade Ltd.||Parallel reed-solomon RAID (RS-RAID) architecture, device, and method|
|US8429341||Apr 23, 2013||Netapp, Inc.||Method and system for responding to file system requests|
|US8539253||Jul 18, 2006||Sep 17, 2013||Netapp, Inc.||System and method for securing information by obscuring contents of a persistent image|
|US8645798||Jan 22, 2013||Feb 4, 2014||Marvell World Trade Ltd.||Parallel Reed-Solomon RAID (RS-RAID) architecture, device, and method|
|US8825970||Apr 26, 2007||Sep 2, 2014||Netapp, Inc.||System and method for mounting a storage volume utilizing a block reference list|
|US9367396 *||Dec 19, 2014||Jun 14, 2016||DSSD, Inc.||Method and system for service-aware parity placement in a storage system|
|US20010039632 *||Jan 25, 2001||Nov 8, 2001||Maclaren John M.||Raid memory|
|US20030051113 *||Oct 11, 2002||Mar 13, 2003||Beardsley Brent Cameron||Method and system for managing meta data|
|US20040034817 *||Aug 15, 2002||Feb 19, 2004||Talagala Nisha D.||Efficient mechanisms for detecting phantom write errors|
|US20040123202 *||Dec 23, 2002||Jun 24, 2004||Talagala Nisha D.||Mechanisms for detecting silent errors in streaming media devices|
|US20040128427 *||Dec 15, 2003||Jul 1, 2004||Kazar Michael L.||Method and system for responding to file system requests|
|US20040133539 *||Dec 23, 2002||Jul 8, 2004||Talagala Nisha D||General techniques for diagnosing data corruptions|
|US20040153746 *||Jul 31, 2003||Aug 5, 2004||Talagala Nisha D.||Mechanisms for embedding and using integrity metadata|
|US20050207049 *||Mar 17, 2005||Sep 22, 2005||Hitachi Global Storage Technologies Netherlands, B.V.||Magnetic disk drive and refresh method|
|US20050278382 *||May 28, 2004||Dec 15, 2005||Network Appliance, Inc.||Method and apparatus for recovery of a current read-write unit of a file system|
|US20050283682 *||Jun 18, 2004||Dec 22, 2005||Hitachi, Ltd.||Method for data protection in disk array systems|
|US20060075281 *||Sep 27, 2004||Apr 6, 2006||Kimmel Jeffrey S||Use of application-level context information to detect corrupted data in a storage system|
|US20060075287 *||Oct 6, 2004||Apr 6, 2006||Vageline Michael P||Detecting data integrity|
|US20060123190 *||Dec 3, 2004||Jun 8, 2006||Weirauch Charles R||System and method for writing data and a time value to an addressable unit of a removable storage medium|
|US20060224823 *||Mar 30, 2005||Oct 5, 2006||Emc Corporation||Sector-edge cache|
|US20060248292 *||Sep 19, 2005||Nov 2, 2006||Tanjore Suresh||Storage processor for handling disparate requests to transmit in a storage appliance|
|US20060294438 *||Aug 29, 2006||Dec 28, 2006||Hitachi, Ltd||Method for data protection in disk array systems|
|US20070027985 *||Aug 1, 2005||Feb 1, 2007||Network Appliance, Inc.||Rule-based performance analysis of storage appliances|
|US20070100783 *||Oct 29, 2005||May 3, 2007||International Business Machines Corporation||Method, system, and program for determining discrepancies between database management systems|
|US20070100910 *||Oct 31, 2005||May 3, 2007||Abhijeet Gole||Circular and bi-directional mirroring of flexible volumes|
|US20070255921 *||Apr 28, 2006||Nov 1, 2007||Abhijeet Gole||Methods of converting traditional volumes into flexible volumes|
|US20080022133 *||Jul 18, 2006||Jan 24, 2008||Network Appliance, Inc.||System and method for securing information by obscuring contents of a persistent image|
|US20080133772 *||Jan 8, 2008||Jun 5, 2008||Kazar Michael L||Method and system for responding to file system requests|
|US20090070345 *||Oct 31, 2008||Mar 12, 2009||Kazar Michael L||Method and apparatus for data storage using striping specification identification|
|US20090132851 *||Nov 20, 2008||May 21, 2009||Arvind Pruthi||Parallel reed-solomon raid (rs-raid) architecture, device, and method|
|US20090271459 *||Oct 29, 2009||Kazar Michael L||Method and system for responding to file system requests|
|US20100114889 *||Oct 30, 2008||May 6, 2010||Netapp, Inc.||Remote volume access and migration via a clustered server namespace|
|US20100235572 *||Mar 13, 2009||Sep 16, 2010||Seagate Technology Llc||Data storage system redundancy scheme verification|
|US20110082974 *||Dec 13, 2010||Apr 7, 2011||Hewlett- Packard Development Company, L.P.|
|US20110202581 *||Aug 18, 2011||Kazar Michael L||Method and system for responding to file system requests|
|CN100437534C||Jul 29, 2005||Nov 26, 2008||惠普开发有限公司;戴尔产品有限合伙人公司;皇家飞利浦电子股份有限公司||Drive tracking system for removable media|
|U.S. Classification||714/6.12, 714/E11.034, 714/800|
|International Classification||G06F11/10, G06F3/06, G06F12/16|
|Cooperative Classification||G06F2211/104, G06F11/1076|
|Feb 24, 1997||AS||Assignment|
Owner name: DATA GENERAL CORPORATION, MASSACHUSETTS
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:PETERSON, GARY S.;REEL/FRAME:008410/0289
Effective date: 19970213
|Aug 29, 2002||FPAY||Fee payment|
Year of fee payment: 4
|Oct 2, 2006||FPAY||Fee payment|
Year of fee payment: 8
|Sep 30, 2010||FPAY||Fee payment|
Year of fee payment: 12